确保病理学家判读一致以保障多中心肿瘤学IVD/CDx临床试验的数据完整性

Background: Multi-centre oncology clinical trials increasingly rely on digital pathology and distributed specialist review, resulting in a heightened risk of inter-pathologist variability. Without robust alignment, differences in tissue interpretation can skew patient inclusion, compromise assay performance, and ultimately threaten the validity of clinical investigations for In Vitro Diagnostic (IVD) and companion diagnostic (CDx) development. Objective: To demonstrate the critical importance of pathologist alignment and the impact of structured interpretation guidance on assay concordance in a multi-reviewer setting. Methods: ARC Laboratories evaluated the effect of pathologist alignment using a gastric cancer cohort (n=20). FFPE samples were stained with a CLDN18 antibody (Ventana CLDN 18 43-14A) on the BenchMark UltraPlus platform. In Phase 1, two independent, CLDN18-trained gastrointestinal pathologists assessed tumor content, tumor positivity, and percentage tumor involvement without a pathology manual or consensus meeting. In Phase 2, both pathologists underwent an alignment process including a pre-read, defined pathology interpretation manual, and agreement meeting before reassessment. Results: In the unaligned phase, pathologists agreed on tumor positivity in only 8/20 cases. Tumor content and percentage tumor interpretation differed by up to 50%, resulting in assay performance that would not meet validation acceptance criteria. Following alignment, concordance increased substantially: 18/20 cases met agreement for tumor positivity, and differences in tumor content estimation decreased to ≤20%. Under these conditions, the assay met all required validation parameters and was deemed suitable for use in clinical trials. Conclusion: This study highlights that even highly trained subspecialist pathologists can exhibit significant variability without structured alignment tools. A defined pathology manual and consensus process markedly improve concordance and are essential for ensuring assay validity in multi-centre or digitally enabled IVD/CDx trials. ARC's findings underscore that systematic pathologist alignment is not optional-it is a foundational requirement for generating reliable, regulatory-ready clinical data.